A broad variety of illnesses involve the JNK family [52, 53]. Indeed, JNKs are thought to be a critical mediator of neuronal response to stress, involving both neuronal survival and death under a variety of conditions [54]. There are at least ten JNK isoforms expressed from three genes, exhibiting differences in substrate and binding protein specificity [6]. Knock-out animal models disclosed different gene product features [16, 9], yet evidence for selective activation of endogenous JNKs is absent. Indeed, although many studies in the literature have addressed the cognitive and molecular consequences of JNK3 ablation in AD, to our knowledge, currently there is no study that analyzes the consequences of JNK3 overexpression on cognitive performance. Thus, the main aim of the present study was to assess the consequences of JNK overexpression, more specifically overexpression of the JNK3 isoform, i.e. the main isoform in the brain.
This work focuses on the EC as it is considered to be one of the key sites for the development of neurodegeneration. The EC is an essential area located in the medial temporal lobe, whose functions include long-term-memory. Interestingly, EC projects to Hp and it receives inputs from other cortical areas. The EC is divided in two main areas: the medial EC (MEC) and the lateral EC (LEC). Both MEC and LEC has shown to have different functional characteristics. The MEC superficial layers comprise several spatially modulated cell types, whereas the LEC's adjacent neurons exhibit only sparse spatial modulation [55–57] and somatosensory information [58–61]. The spatial information coming from the MEC together with the non-spatial information processed from the LEC are integrated in the EC [62–65]. EC is one of the earliest affected areas in neurodegenerative disorders such as AD, indicating the essential participation of EC in cognition [66]. Although the reason behind this early EC impairment in AD is still unknown, a specific vulnerability to aging and AD of the EC neurons is hypothesized [67], that induces a significant neuronal death in this area during the first stages of the disease [68]. Noteworthy, amyloid protein and hyperphosphorylated Tau aggregation, i.e. the main AD histopathological characteristics, appear first in the EC in mild AD and are not disseminated to other areas such as the Hp until more advance stages of the disease [69]. Hence, it has been suggested that the neurodegeneration that starts in EC neurons is transferred to the Hp, inducing the disruption of the cortical-hippocampal network in AD patients. In light of these important findings, in this study it was decided to induce JNK3 overexpression in both MEC and LEC, in order to elucidate if increased levels of JNK3 could lead to a cortical-hippocampal network dysfunction and ultimately to cognitive alterations. Furthermore, JNK3 overexpression was induced in wild type mice, in an attempt to mimic early stages of AD when amyloid plaque or neurofibrillary tangle accumulations are still absent.
Our results showed that although viral infection was conducted in EC (MEC and LEC), JNK3 overexpression is also observed in the Hp, concluding that changes in the EC can lead directly to downstream modifications in its main afferent areas, such as the Hp, leading to aberrant network activity as it has been observed in mouse models and human AD patients [70, 71]. More importantly, we demonstrated that JNK3 overexpression was associated with a behavioral impairment of associative memory, assessed by the NORT. A significant role in object recognition and novelty detection has already been assigned to the EC [72]. In particular, information from EC can be acquired in the Hp through the complex integration of spatial information coming from MEC with non-spatial input from the LEC [64, 65]. Specifically, a population of LEC cells have been identified, some of them firing at the objects and other cells firing at places where objects were located on previous trials [72]. In addition, LEC is needed to recognize items encountered in a particular context [73] and the specific lesion of the LEC impairs the capacity to discriminate either novel object-place or novel object-place-context associations [73]. Therefore, in light of our results, it seems that the induction of JNK3 overexpression in the EC affects the integration of information in the Hp, leading to cognitive deficiencies. On the contrary, no alterations were observed in the MWM task after JNK3 overexpression. The MWM is a classical test to assess spatial and thus hippocampus-dependent memory performance [74]. Therefore, our results suggest that the increase of JNK3 obtained in the Hp is not strong enough to induce a spatial learning impairment, as it occurs in early stages of AD. Probably, a higher JNK3 dissemination in the hippocampus is necessary to induce deficits in spatial learning
The proof that JNK accumulation is associated with inflammatory pathway activation [75] raises the main question of whether brain neuroinflammation is involved in the early behavioral deficits found in the present study after JNK3 overexpression induction. Inflammation is the first reaction from our body's immune system to pathogens or irritation and it is a two-edged sword. It protects tissue against invading agents under acute circumstances and encourages healing. On the other hand, it can cause severe damage to the host's own tissue if it is chronically maintained. While the CNS is recognized as an immune-privileged organ, there is growing evidence that inflammation is directly involved in the pathogenesis of a number of neurodegenerative diseases, including AD, multiple sclerosis (MS), and HIV-associated dementia [76, 77, 78]. Chronic inflammation-mediated tissue injury can be particularly damaging to the brain, as neurons are usually irreplaceable. In particular, it has been extensively demonstrated the involvement of astrocytes and microglia in the pathological process of AD. Indeed, it has been observed in AD animal models and patients that the cognitive deficiencies are accompanied by chronic glial activation and pro-inflammatory cytokine production [79]. Consequently, pathological markers indicative of astrogliosis and microgliosis are correlated with cognitive disturbances in AD [80, 81, 82]. Increased levels of pro-inflammatory cytokines are detected in early phases of clinical AD patients and it is suggested that those cytokines contribute to the neurotoxicity observed in AD late stages [83, 84, 85, 86]. In agreement with those studies, our data demonstrated that overexpression of JNK3 induced all the pathological markers observed in early-stages of AD brains, i.e., microgliosis, astrogliosis and pro-inflammatory cytokine (IL-1β, IL-6, TNFα) release that could contribute to the cognitive deficiencies observed in the JNK3-induced mice. Interestingly, although all those markers were strongly increased in the EC (the injection area), neuroinflammation was milder in the Hp. This could also explain the absence of cognitive alterations in the MWM.
Apart of its central role in neuroinflammation, JNK kinase can participate in AD pathology by its implication in Tau phosphorylation and subsequent neurofibrillary tangles formation [87]. It has been demonstrated by in vitro experiments that JNK3 isoform can be autophosphorylated and then, it can contribute to Tau hyperphosphorylation [88]. Tau hyperphosphorylation induces its aberrant misfolding, following by its dissociation from microtubules and aggregation in neurofibrillary tangles. In order to study the implication of JNK3 on the conformation of Tau aberrant misfolding, two different conformations were studied: ALZ50 and MC1. ALZ-50 has been detected in brain homogenates [89] inside susceptible neurons [43, 89–92]. MC1 appeared to be a good marker for early aggregation of Tau protein, before the appearance of neurofibrillary tangles [93–95]. Another modification associated with Tau deposition in AD is truncation [96, 97]. Several authors consider that Tau truncation in the C-terminus precedes Tau assembly in paired helical filaments [49, 50, 96–99] and truncation has been associated with early as well as late stages of AD pathology [100, 101, 102, 103]. Interestingly, Tau truncation is frequently preceded by Tau Ser422 phosphorylation. In our hands, all the aberrant conformations studied (ALZ50, MC1, truncated Asp421 Tau and Tau Ser422) appeared to be strongly increased after JNK3 overexpression, suggesting that Tau misfolding and subsequent microtubule disaggregation could be also underlying the cognitive deficiencies observed in AAV-JNK3 mice. Noteworthy, the fact that in the Hp Tau misfolding assessment did not reach statistical significance might ground the lack of cognitive impairment in the MWM task.
In summary, the data obtained in the present study indicate that activation of inflammatory signals and induction of Tau in vivo misfolding triggered by an enriched JNK3 environment is a significant early event during the progressive EC dysfunction. Therefore, JNK3 overexpression can lead to the triggering of cognitive dysfunction resulting in the dissemination of neurodegeneration from EC to Hp and may be at the origin of the changes observed in early stages of AD.